Evaluation of the Glasgow-Blatchford score in predicting clinical outcomes in upper gastrointestinal bleeding

Abstract

Introduction and aim. Acute upper gastrointestinal bleeding is a common cause of emergency admissions with potentially serious outcomes. Early evaluation of patients is crucial to predict morbidity, recurrence of bleeding, and mortality. The Glasgow Blatchford score (GBS) is a validated scoring system used to predict the need for medical interventions such as blood transfusion, endoscopy, and surgery. This study aimed to explore the correlation of GBS with prognostic markers in patients with upper gastrointestinal bleeding.

Material and methods. This retrospective study included patients >18 years old admitted to Hitit University Corum Erol Olcok Training and Research Hospital due to upper gastrointestinal bleeding between December 2022 and May 2023. Exclusion criteria were insufficient endoscopy or data or pregnancy. GBS scores were calculated at the initial presentation for each patient and their association with prognostic markers and mortality was analyzed. Comparison of numerical measurements between independent groups was evaluated using the Mann-Whitney U test and categorical variables were evaluated using the Chisquare test. Spearman coefficients were used for correlations. ROC analysis was used to determine the sensitivity and specificity of GBS to predict endpoints. The predictive factors for the endpoints were investigated using logistic regression analysis.

Results. A total of 140 patients were enrolled in the study. GBS was significant in predicting the need for blood transfusion (OR: 1.493, 95% CI: 1.297–1.719, p<0.001), need for endoscopic intervention (OR: 1.248, 95% CI: 1.089–1.430, p=0.001), and preference for ward/intensive care unit (OR: 0.869, 95% CI: 0.790–0.953, p=0.003). For predicting mortality, Charlson Comorbidity Index (OR: 1.023, CI=1.008–1.437, p=0.046) was significant. GBS was not significant for predicting mortality (p=0.582). The area under the curve (AUC) of GBS with a cut-off of 9.5 for mortality was 0.64 (95% CI 0.513–0.775, p=0.032) with a sensitivity of 68.2% and specificity of 52.5%, AUC 0.752 (95% CI 0.653–0.851, p<0.001) for the need for endoscopic intervention with a sensitivity of 90% and specificity of 50.8%, AUC 0.729 (95% CI 0.646–0.812, p<0.001) for admission to intensive care with a sensitivity of 70.1% and specificity of 58.9% and AUC 0.853 (95% CI 0.782–0.924, p<0.001) for the need for blood transfusion with a cut-off of 8.5 with a sensitivity of 84.9% and specificity of 75.5% for the selected.

Conclusion. The GBS did not predict mortality, but effectively predicted the need for blood transfusion, endoscopic intervention, and intensive care unit admission. The Charlson comorbidity index was predictive for mortality in this study group.